Usual first-line treatments for major depressive disorder (MDD) show less than a 40% remission rate. For many patients, the wrong treatment has significant individual and societal costs due to continued distress, risk of suicide, loss of productivity, wasted resources and potential neural effects associated with 2-3 months of an ineffective strategy. Though treatments are highly effective in some individuals, there is no reliable way to match patients to their best option. Therefore, a long term goal is to develop a clinically viable algorithm that selects the best treatment and avoids ineffective treatments, whil also identifying patients that require alternatives to standard first-line options. The objective o this proposal is to test the efficacy of a novel imaging biomarker developed to stratify patients into two subtypes that predict the likelihood of remission to monotherapy with cognitive behavioral therapy (CBT) or escitalopram (sCIT)--two standard first-line treatments for MDD. Our central hypothesis is that use of insula metabolism as a treatment selection biomarker (TSB) to assign treatment will increase remission rates to at least 50%, exceeding the usual 35-40% commonly reported with usual care. Data from the previous funding period identified the insula as the best discriminator of remission/non response to CBT and sCIT. The data further indicated other potential biomarkers that identify patients who will fail to remit to combined treatment with both CBT and sCIT. We will examine two specific aims: 1) To prospectively test the efficacy of the insula TSB to assign individual MDD patients to treatment with either CBT or sCIT; and 2) To further characterize brain subtypes predictive of combined treatment failure., Treatment assignment will be determined by the level of glucose metabolism in the right anterior insula measured with positron emission tomography (FDG PET). We will derive each patient's insula TSB using a simple region/whole brain ratio approach and a fixed cut-off. Patients will be assigned to a 12-week treatment course of CBT or sCIT using the insula TSB value. Patients who do not remit to their TSB-assigned first treatment will receive a second 12-week course of combined CBT and sCIT. Baseline whole brain metabolism will be evaluated in patients who do not remit after both monotherapy and combination therapy to determine biomarkers of dual failure. Lastly, we will examine regional metabolic changes over the first 12 weeks of treatment to characterize mechanisms mediating remission across insula subtypes. This proposal is innovative because it tests a novel imaging biomarker strategy developed specifically to prospectively stratify patients into two distinct treatment specific subtypes that will predict the likelihood of response to two standard first line therapies for a major depressive episode. In addition, this study will further characterize a second biomarker that identifies patients who fail to remit to both treatments. The proposed research is significant because it directly addresses the need for evidence-based methods for selecting optimal treatments for individual depressed patients with the ultimate goal of improving treatment outcomes.